Use ICU by default at initdb time.

If the ICU locale is not specified, initialize the default collator
and retrieve the locale name from that.

Discussion: https://postgr.es/m/510d284759f6e943ce15096167760b2edcb2e700.camel@j-davis.com
Reviewed-by: Peter Eisentraut

6 files changed, 52 insertions, 3 deletions

diff --git a/contrib/citext/expected/citext_utf8.out b/contrib/citext/expected/citext_utf8.out
index 666b07ccec4..77b4586d8fa 100644
--- a/contrib/citext/expected/citext_utf8.out
+++ b/contrib/citext/expected/citext_utf8.out
@@ -1,9 +1,16 @@
 /*
  * This test must be run in a database with UTF-8 encoding
  * and a Unicode-aware locale.
+ *
+ * Also disable this file for ICU, because the test for the the
+ * Turkish dotted I is not correct for many ICU locales. citext always
+ * uses the default collation, so it's not easy to restrict the test
+ * to the "tr-TR-x-icu" collation where it will succeed.
  */
 SELECT getdatabaseencoding() <> 'UTF8' OR
-       current_setting('lc_ctype') = 'C'
+       current_setting('lc_ctype') = 'C' OR
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
        AS skip_test \gset
 \if :skip_test
 \quit
diff --git a/contrib/citext/expected/citext_utf8_1.out b/contrib/citext/expected/citext_utf8_1.out
index 433e9853497..d1e1fe1a9d8 100644
--- a/contrib/citext/expected/citext_utf8_1.out
+++ b/contrib/citext/expected/citext_utf8_1.out
@@ -1,9 +1,16 @@
 /*
  * This test must be run in a database with UTF-8 encoding
  * and a Unicode-aware locale.
+ *
+ * Also disable this file for ICU, because the test for the the
+ * Turkish dotted I is not correct for many ICU locales. citext always
+ * uses the default collation, so it's not easy to restrict the test
+ * to the "tr-TR-x-icu" collation where it will succeed.
  */
 SELECT getdatabaseencoding() <> 'UTF8' OR
-       current_setting('lc_ctype') = 'C'
+       current_setting('lc_ctype') = 'C' OR
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
        AS skip_test \gset
 \if :skip_test
 \quit
diff --git a/contrib/citext/sql/citext_utf8.sql b/contrib/citext/sql/citext_utf8.sql
index d068000b423..8530c68dd7e 100644
--- a/contrib/citext/sql/citext_utf8.sql
+++ b/contrib/citext/sql/citext_utf8.sql
@@ -1,10 +1,17 @@
 /*
  * This test must be run in a database with UTF-8 encoding
  * and a Unicode-aware locale.
+ *
+ * Also disable this file for ICU, because the test for the the
+ * Turkish dotted I is not correct for many ICU locales. citext always
+ * uses the default collation, so it's not easy to restrict the test
+ * to the "tr-TR-x-icu" collation where it will succeed.
  */
 
 SELECT getdatabaseencoding() <> 'UTF8' OR
-       current_setting('lc_ctype') = 'C'
+       current_setting('lc_ctype') = 'C' OR
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
        AS skip_test \gset
 \if :skip_test
 \quit
diff --git a/contrib/unaccent/expected/unaccent.out b/contrib/unaccent/expected/unaccent.out
index ee0ac71a1cc..cef98ee60cc 100644
--- a/contrib/unaccent/expected/unaccent.out
+++ b/contrib/unaccent/expected/unaccent.out
@@ -1,3 +1,12 @@
+-- unaccent is broken if the default collation is provided by ICU and
+-- LC_CTYPE=C
+SELECT current_setting('lc_ctype') = 'C' AND
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
+	AS skip_test \gset
+\if :skip_test
+\quit
+\endif
 CREATE EXTENSION unaccent;
 -- must have a UTF8 database
 SELECT getdatabaseencoding();
diff --git a/contrib/unaccent/expected/unaccent_1.out b/contrib/unaccent/expected/unaccent_1.out
new file mode 100644
index 00000000000..0a4a3838abd
--- /dev/null
+++ b/contrib/unaccent/expected/unaccent_1.out
@@ -0,0 +1,8 @@
+-- unaccent is broken if the default collation is provided by ICU and
+-- LC_CTYPE=C
+SELECT current_setting('lc_ctype') = 'C' AND
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
+	AS skip_test \gset
+\if :skip_test
+\quit
diff --git a/contrib/unaccent/sql/unaccent.sql b/contrib/unaccent/sql/unaccent.sql
index 3fc0c706be3..027dfb964a7 100644
--- a/contrib/unaccent/sql/unaccent.sql
+++ b/contrib/unaccent/sql/unaccent.sql
@@ -1,3 +1,14 @@
+
+-- unaccent is broken if the default collation is provided by ICU and
+-- LC_CTYPE=C
+SELECT current_setting('lc_ctype') = 'C' AND
+       (SELECT datlocprovider='i' FROM pg_database
+        WHERE datname=current_database())
+	AS skip_test \gset
+\if :skip_test
+\quit
+\endif
+
 CREATE EXTENSION unaccent;
 
 -- must have a UTF8 database